[论文解读] StoryLensEdu: Personalized Learning Report Generation through Narrative-Driven Multi-Agent Systems

Leixian Shen, Yan Luo+5 more|arXiv (Cornell University)|Feb 19, 2026

Intelligent Tutoring Systems and Adaptive Learning被引用 0

一句话总结

tldr: StoryLensEdu introduces a narrative-driven, three-agent system that generates personalized, engaging learning reports with interactive question-answering for self-regulated learning.

ABSTRACT

Personalized feedback plays an important role in self-regulated learning (SRL), helping students track progress and refine their strategies. However, current common solutions, such as text-based reports or learning analytics dashboards, often suffer from poor interpretability, monotonous presentation, and limited explainability. To overcome these challenges, we present StoryLensEdu, a narrative-driven multi-agent system that automatically generates intuitive, engaging, and interactive learning reports. StoryLensEdu integrates three agents: a Data Analyst that extracts data insights based on a learning objective centered structure, a Teacher that ensures educational relevance and offers actionable suggestions, and a Storyteller that organizes these insights using the Heroes Journey narrative framework. StoryLensEdu supports post-generation interactive question answering to improve explainability and user engagement. We conducted a formative study in a real high school and iteratively developed StoryLensEdu in collaboration with an e-learning team to inform our design. Evaluation with real users shows that StoryLensEdu enhances engagement and promotes a deeper understanding of the learning process.

研究动机与目标

• Objective 1: Address cognitive overload and limited interpretability of traditional feedback by creating engaging narrative reports.
• Objective 2: Automate context-aware extraction of learning insights aligned to a learning objectives graph.
• Objective 3: Coordinate specialized agents to produce pedagogically meaningful feedback and coherent storytelling.
• Objective 4: Enable post-generation interactive exploration to improve explainability and learner engagement.
• Objective 5: Evaluate the system in real-world settings to assess impact on engagement and understanding of learning progress.

提出的方法

• Method 1: Introduce a learning objective-centered data structure modeled as a directed graph to map student performance to objectives.
• Method 2: Use a three-agent report generation engine (Data Analyst, Teacher, Storyteller) powered by LLMs to extract insights, reason pedagogically, and narrate the report.
• Method 3: Embed the Hero’s Journey narrative framework within the Storyteller agent to structure the report.
• Method 4: Provide an interaction module for post-generation question answering grounded in the objective graph.
• Method 5: Implement a Q&A model that maps user selections to relevant objectives and retrieves data via targeted SQL queries for context-aware responses.
• Method 6: Adopt a metadata-first workflow where the Data Analyst outputs deterministic JSON insights that the Teacher converts into natural language explanations.

实验结果

研究问题

• RQ1 研究问题 1：如何使个性化反馈在传统文本报告和仪表板之外更易于解释和更具吸引力？
• RQ2 研究问题 2：以叙事驱动的多代理系统能否有效地使用学习目标图诊断和支持自我调节学习？
• RQ3 研究问题 3：交互式后生成提问是否提升了对学习报告的可解释性和学习者参与度？
• RQ4 研究问题 4：英豪旅程叙事在组织教育反馈以提升动机与清晰度方面的作用是什么？

主要发现

• 关键发现 1：系统将自动洞察提取与叙事框架和交互式问答相结合。
• 关键发现 2：与真实用户的评估显示参与度提高，对学习过程的理解更深。
• 关键发现 3：学习目标图实现了情境感知的反馈，并可追溯地对齐课程目标。
• 关键发现 4：数据分析师提供多维时间序列诊断，教师提供经教育学依据的建议。
• 关键发现 5：讲述者利用英豪旅程将洞察整合为一个连贯的叙事以提升动机和理解力。